The overall objective of the proposed research is to evaluate the properties of the technique of measuring delinquent behavior by the self report method in order to recommend the most reliable and valid instruments and procedures. To this end a sample of the general juvenile population -- stratified by race, sex, SES, and delinquency status -- will be tested using various self-report methods. One specific aim that contributes to this objective is the use of a reverse record check among juveniles with known court contacts. This procedure will reveal information about those who do and do not report court contacts and known offenses when various self-report procedures are used. Furthermore, it will be possible to isolate whether court contacts for specific kinds of offenses are more often reported and whether recent contacts are more accurately reported. The second specific objective is to examine the differential reliability and validity (across respondent characteristics such as race, sex, SES, and official delinquency status) of the various self-report procedures included in the research design. The vaidity of the responses will be assessed primarily by the known court contacts and known offenses of the respondents and by the analysis of interview using the Psychological Stress Evaluator. In addition, validity will be assessed by means of the "nominated group" approach and by "deep probe" interviews. Reliability will be assessed using both test/re-test and equivalent forms procedures. The Warner randomized response (Warner, 1965; Greenberg, et al., 1969) approach will also be studied to determine whether this approach, which offers added security to the subject, results in higher levels of reported delinquent involvement than standard techniques.